Skip to content
TrustList
Launches

Launches

The best new software, tools and AI models - ranked by the community. Every launch links to a full TrustList profile with real, verified reviews.

By field:Modern carsRobotics & automationAutonomous vehicles & mobility

Showing every launch, newest first. Pick a range above to rank it by community votes - weighted by each voter's trust, with editors able to reorder a launch explicitly. Featuring does not change this order: it only decides what the Featured tab shows, and that tab includes TrustList's own products.

No account yet? Email hello@trustlist.uk with launch in the subject line and we'll set it up for you - tell us the name, a one-line tagline, what it does, the date, and a link.

Sun, 22 Nov 2026

Showing 61-120 of 229 launches

  • 61

    Google’s open multimodal embedding model, released 6 October 2026 under Apache 2.0: 740M parameters on the Gemma 4 architecture, one vector space for text, images, audio and video, 8K context, vectors truncatable from 768 dimensions.

    TL score
  • 62
    KolibriAI model

    Aleph Alpha’s open-weight English-German mixture-of-experts model, released 3 October 2026 under Apache 2.0: 78B parameters in total with 3B active, and a context of up to 1M tokens, for self-hosting.

    TL score
  • 63
    AstaBrief 8BAI model

    Ai2’s open-weight 8B model for cited scientific reports, released 2 October 2026: it turns a research question and retrieved literature into a report with citations. Weights and training data published; runs as Fast mode in Asta.

    TL score
  • 64

    An open-weight translation model from Bilibili’s Index team, published 2 October 2026 under Apache 2.0: a 35B mixture-of-experts model with 3B active parameters that translates across 150 languages and follows style and terminology instructions.

    TL score
  • 65
    ClefAI model

    Cloudflare’s open-weight decision model, released 1 October 2026 under Apache 2.0: classifies text or images into typed answers with probabilities so agents can act without a human check. 64K context, on Workers AI.

    TL score
  • 66
    Clef-flashAI model

    The latency-optimised member of Cloudflare’s Clef decision-model family, released 1 October 2026 under Apache 2.0: structured, probability-scored classification of text or images for agents, built on a 9B backbone.

    TL score
  • 67

    An open-source 2B “decision model” from AWS’s Strands Agents project, released 1 October 2026 under Apache 2.0: it picks between given options or returns a score with a reliability figure, fast enough for a local CPU.

    TL score
  • 68

    Microsoft AI’s real-time speech-to-text model, released 1 October 2026: streaming transcripts in 60 languages with continuous language detection and first partial results in just over 100 ms, at $0.54 per audio hour.

    TL score
  • 69

    Microsoft AI’s multilingual text-to-speech model, released 1 October 2026: 23 languages and 26 locales, one voice across all of them with a native accent, and voice cloning from seconds of audio, at $22 per million characters.

    TL score
  • 70

    The high-volume variant of Microsoft AI’s MAI-Voice-2.1, released 1 October 2026: the same 23 languages and voice cloning, with 150 ms end-to-end latency for 45 seconds of audio, at $15 per million characters.

    TL score
  • 71
    FLUX 3 ImageAI model

    Black Forest Labs’ image generation and editing model, released 1 October 2026: bounding-box layouts, edits that leave untouched pixels alone, multiple reference images and native 2K and 4K output, via API or commercial weights.

    TL score
  • 72

    Perplexity’s open-weight “decision model”, published 1 October 2026 under Apache 2.0: a 27B fine-tune of Qwen3.8-27B that classifies and chooses among given options rather than writing free text, for routing and checks in AI systems.

    TL score
  • 73

    Google’s Gemini 4 frontier model, announced on 30 September 2026 and so far released only to trusted cyber defenders in its Fairwind Program; $2 in and $10 out per million tokens at launch, then $4 and $20.

    TL score
  • 74
    Embed 5 ProAI model

    Cohere’s multimodal embedding model for enterprise search and retrieval, released 30 September 2026: text, images, mixed and parsed documents in 100+ languages, 128K context, $0.12 per million text tokens.

    TL score
  • 75
    Embed 5 FastAI model

    The lower-cost model of Cohere’s Embed 5 family, released 30 September 2026: multimodal, multilingual embeddings with 128K context and the same dimension and quantisation options as Embed 5 Pro, at $0.08 per million text tokens.

    TL score
  • 76
    apex-flash-1AI model

    Cantina Security’s first open-weight security model, published 30 September 2026 under MIT: a reinforcement-learning post-train of GLM-5.3-Flash for vulnerability research, reading code, using tools and verifying exploits in a target.

    TL score
  • 77
    GPT-6.1 SolAI model

    OpenAI’s mid-tier GPT-6 model at $2 in and $10 out per million tokens, a fifth of GPT-6 Astra’s price, with a 1.05M-token context, 128K output tokens, text and image input and multi-agent support in beta.

    TL score
  • 78

    Inception’s diffusion language model for real-time voice agents, released 29 September 2026 for enterprise customers: 128K input and 50K output tokens, $0.40 in and $1.50 out per million tokens before a 50% launch discount.

    TL score
  • 79

    VIDRAFT’s open 180B reasoning model: Alibaba’s Qwen3.8-Flash-Next retrained on its own checked answers, reading text and images, with a built-in confidence score that is returned before it answers.

    TL score
  • 80

    Anthropic’s mid-tier model at $2 in and $10 out per million tokens, half Opus 5.5’s price, with a 1M-token context, 128,000 output tokens and a retirement floor of 28 September 2027. It replaces Sonnet 5.

    TL score
  • 81
    IQuest-Q1AI model

    IQuestLab’s open agentic-coding model: a 320B mixture of experts with 15B active, a 524,288-token context and English and Chinese text, for self-hosting with SGLang or vLLM. Modified MIT licence with a display requirement.

    TL score
  • 82
    Eleven v4AI model

    ElevenLabs’ text-to-speech model released 28 September 2026: a new architecture for expressive speech in 90+ languages, with multiple speakers, inline emotion and sound-effect tags, and long-form stitching. Hosted only.

    TL score
  • 83

    The low-latency member of ElevenLabs’ v4 text-to-speech family, released 28 September 2026 for voice agents: ElevenLabs states a median inference latency of about 100 ms and about 150 ms to first speech. Hosted only.

    TL score
  • 84
    Drex v1.5AI model

    Nace AI's decision model, about 9B parameters: takes a state as text or JSON plus typed questions and returns a probability for each option, not free text. Open weights under a modified Open RAIL-M licence.

    TL score
  • 85

    NaiveAI’s open MIT-licensed coding and research model: a 309B mixture of experts with 15.5B active and a native 1M-token context, with an announced API price of $0.10 in and $0.40 out per million tokens.

    TL score
  • 86

    InternLM’s open decision models in 0.8B, 2B and 4B sizes: given a state, optional images and typed questions, they return a probability for each allowed answer in one pass instead of writing text. Apache 2.0.

    TL score
  • 87

    Perceptron’s multimodal model for robots, drones, smart glasses and phones: it takes text, images, video and audio and returns text plus points, boxes, polygons, clips and object tracks, at $0.15 in and $1.50 out per million tokens.

    TL score
  • 88

    Fastino’s open Apache-2.0 decision classifier that runs on a CPU: give it any set of labels at call time and it returns typed decisions with confidence scores for routing, triage, intent and moderation.

    TL score
  • 89
    LevAI model

    Interfaze’s open Apache-2.0 decision model: a 4B adapter that answers yes/no, pick-one and score questions about a ticket, email or JSON with probabilities instead of text, and accepts code written for TypeSafe’s Jev API.

    TL score
  • 90

    Google’s expressive text-to-speech model: 130 languages, voice design from a written description, $0.50 in and $9.00 out per million tokens until 31 December 2026 — and double that from 1 January 2027.

    TL score
  • 91

    The cheaper Gemini 3.8 speech model for dubbing and voice agents: 101 languages, $0.50 in and $6.00 out per million tokens until 31 December 2026, doubling from 1 January 2027. Replaces the 3.1 Flash TTS preview.

    TL score
  • 92

    The speed and cost tier of Recraft’s V4.1 text-to-image family: $0.007 an image through the API, against $0.035 for standard V4.1, with a median of 1.3 seconds per image on Recraft’s own tests.

    TL score
  • 93

    Black Forest Labs’ 7B open-weight model that turns camera views, joint positions and a plain-language instruction into robot-arm actions. Commercial use is free only below US$5 million in annual revenue.

    TL score
  • 94

    NVIDIA’s open-weight 100M speaker-diarization model: it labels who spoke when for up to eight speakers, streaming or offline, and ranked first of twelve systems on Voice Arena’s independent early results.

    TL score
  • 95
    Ember-1AI model

    Fireworks AI’s retrain of Kimi K3 to reason with fewer tokens: about 40% fewer on the developer’s own tests at similar quality, $3 in and $15 out per million tokens, as a two-week research preview.

    TL score
  • 96

    Together AI’s experimental open 4B decision model, trained to pick one option from a list the way TypeSafe’s Jev does, with the full recipe published. The licence for the weights is still being finalised.

    TL score
  • 97
    CLM-v0.1-8BAI model

    An open Apache-2.0 “contrastive language model” that scores candidate actions against a situation, so an agent can pick or check its next step without generating text. The authors report up to 9x lower latency than Jev.

    TL score
  • 98
    PianissimoAI model

    An open 0.6B Swedish speech-to-text model from Klang AI, trained on about 50,000 hours of Swedish, with word error rates near 4.5% on the developer’s own tests and a CC BY 4.0 licence that allows commercial use.

    TL score
  • 99

    Xiaomi's open-weight flagship: 1.02T parameters with 42B active, a 1M-token context and text, image, video and audio input, at $0.435 in and $0.87 out per million tokens — the price of the mimo-v2.5-pro that Xiaomi retires on 21 October.

    TL score
  • 100

    The smaller MiMo-V2.6: 309B parameters with 15B active, the same 1M context and omni-modal input, open under MIT, at $0.14 in and $0.28 out per million tokens — and the replacement for mimo-v2.5 before 21 October.

    TL score
  • 101
    Solar Mini 4AI model

    Upstage's compact language model for agentic work, with 35B parameters (3B active), a 512K-token context and Korean, English and Japanese support, priced at $0.10 in and $0.40 out per million tokens.

    TL score
  • 102
    GPT-6 SolAI model

    The working tier of the GPT-6 family at half its predecessor’s price: $2 in and $10 out per million tokens, $0.20 cached, a 1,050,000-token context, 128,000 output tokens and an April 2026 knowledge cutoff.

    TL score
  • 103
    GPT-6 LunaAI model

    The cheapest GPT-6: $0.10 in and $0.50 out per million tokens, $0.01 cached, with the same 1,050,000-token context and 128,000-token output limit as GPT-6 Sol and a May 2026 knowledge cutoff.

    TL score
  • 104

    Anthropic’s new default model at $4 in and $20 out per million tokens, a 1M-token context, 128,000 output tokens, a June 2026 cutoff — and a published retirement floor of 22 September 2027, a year to the day from release.

    TL score
  • 105

    An open-weight 6B text-to-image model from Ant Group’s inclusionAI lab for interface screens, dashboards, infographics and posters with legible text and transparent backgrounds, under the MIT licence.

    TL score
  • 106

    An open-weight 6B model from Ant Group’s inclusionAI that splits a finished, flattened design back into separate transparent layers that can be edited, under the MIT licence.

    TL score
  • 107

    BottleCap AI’s fine-tune of Qwen3.8-27B that thinks less: 37% fewer reasoning tokens for under one point of accuracy on the developer’s tests. Free for small businesses; larger ones need a commercial licence.

    TL score
  • 108
    Grok 4.7AI model

    The new Grok flagship, at Grok 4.6's price: $2 in and $6 out per million tokens under 200k-token prompts, a 500k context, reasoning from low to xhigh — and, on the Responses API, reasoning returned encrypted whether asked for or not.

    TL score
  • 109
    Altar-1AI model

    Aikido Security's open-weight security model, a pruned 4-bit version of GLM-5.3 that keeps 168 of 256 experts in 328 GB, released 21 September 2026 to run on-premises on four H200 GPUs under the GLM-5.3 licence.

    TL score
  • 110

    Paradigma's first model, a 1B-parameter dense transformer trained from scratch for competition-level mathematics, handling sequences of about 131K tokens, with Apache 2.0 weights; it is lightly instruction-tuned and not a general assistant.

    TL score
  • 111

    Light Origins' open-weight research preview that turns text instructions into whole-body humanoid motion, built on Qwen3.5-4B and released under Apache 2.0 on 21 September 2026; a separate controller is needed to drive a robot.

    TL score
  • 112

    A 600B mixture-of-experts flagship with 27B active per token and a million-token window, at $1.00 per million in and $2.70 out — priced well under the frontier it is measured against.

    TL score
  • 113

    A 7B image model that generates and edits in one checkpoint, with native transparent output — and, unlike every Qwen image release before it, a licence that forbids commercial use.

    TL score
  • 114

    Alibaba's full-duplex speech-to-speech model on Model Studio, released 20 September 2026, with a 262,144-token context, function calling, web search and voice cloning; audio output costs $24 per million tokens in Singapore.

    TL score
  • 115
    Hemmingway-1AI model

    A 27B open-weight fine-tune of Qwen3.8-27B for emails, messages and short stories, with a 262,144-token context; weights under CC BY-NC 4.0 since 22 September, hosted API $0.24 in and $0.90 out per million tokens.

    TL score
  • 116
    KevAI model

    Jared Palmer's open-source family of small decision models (0.8B, 4B and 9B) on Qwen3.5 that return typed yes/no, choice and score probabilities, released 20 September 2026 under Apache 2.0 with training code and evaluation data.

    TL score
  • 117

    CloudRaker's hosted decision model behind Milliseconds.ai, launched 19 September 2026, returning typed labels, yes/no answers, scores and extracted fields from text or images for $0.04 per million input tokens, with output free.

    TL score
  • 118

    Takes text, images, audio and video in one request across a million tokens of context — up to an hour of continuous audio-video per call, at $0.15 per million input tokens.

    TL score
  • 119
    Needle 3AI model

    An 8-29 MB automation model that runs tool calls and structured extraction on a phone, a wearable or a microcontroller, with no API bill at all.

    TL score
  • 120

    The same GLM-5.3-Flash served faster: 200 tokens a second, 2.5x the price of the model it is built from, and no claim of extra capability anywhere in the vendor's own documentation.

    TL score